A rule feels trustworthy the moment it’s written down — the writing makes it look finished. This activity separates the two things that are actually independent: whether a rule is well-reasoned, and whether the data it was built from looked like the world it gets applied to. Kids can feel the gap between those two the moment they count their misses.
A rule that explains every example in front of you can feel like the best possible rule. It is often just the rule most tightly stitched to the examples you happened to have. This tournament makes that distinction visible and scored, which is a more durable lesson than being told about it.
The opening line of the Academy’s mission is that AI is already sorting classrooms, workplaces, and public life, and that it rarely explains itself. A ledger like this turns that claim into something a student can check for themselves in a single day, in their own life, without needing to trust anyone else’s account of it.
A system doesn’t get to be uncertain about what “late” means — it needs one exact rule to check against, every time. Long before any of that gets written as code, a person has to make the fuzzy word precise, and that step is where a value judgment quietly gets buried. This activity makes students do that step themselves, by hand, so the judgment call is visible instead of hidden inside a system.
Asking “why” of a system that won’t answer directly is a specific, learnable skill — not a personality trait. Twenty questions is a tight enough budget that students have to think strategically about which question narrows the possibilities fastest, which is exactly the discipline real oversight of an opaque system requires.
Most of what a rule-based system does happens somewhere nobody in the room can watch. The realistic response isn’t to give up on understanding it — it’s to test it carefully and build a model of its behavior from the outside. That’s an inductive-reasoning skill, and it’s exactly what this game rehearses, with nothing higher-tech than yes and no.
Kids are taught that fairness has one definition, and that disagreements about it are just someone being unreasonable. This activity shows something closer to the truth: equal treatment, need-based priority, and first-come-first-served are all genuinely fair principles, and they routinely point in different directions on the same decision. Naming that disagreement — rather than resolving it — is the actual skill.
A rule almost never announces who it will exclude — that only becomes visible once real cases pass through it. Writing the rule before seeing the people it will sort, the way this activity forces, is close to how many real systems actually get built: the categories come first, the consequences come later, and by the time the consequences are visible the rule already feels settled.
Most lessons about questioning automated decisions stay abstract. A story makes it concrete: to write an ending where someone actually asks a question, students have to invent the specific thing that made asking possible — knowing who to ask, having evidence, or simply being allowed to speak up. That detail is the real lesson, and it only shows up if they have to write it, not just discuss it.
Most decisions about automated tools get made in a room that doesn’t include everyone the decision affects. This circle puts every one of those missing voices in the room on purpose, including the ones with the least power. Feeling whose concern gets the least airtime — because the format runs out of time before it runs out of speakers — teaches something a summary of “stakeholders” cannot.